Iron Metabolism in Rheumatoid Arthritis By F. DOUGLAS RAYMOND, MORRISA. BOWIEAND ANN DUGAN Patients with rheumatoid arthritis were found to have a rapid plasma clearance of iron, a low serum iron, and a diminished absorption of iron from the gastrointestinal tract. These abnormalities appear to be manifestations of an abnormal iron metabolism which is probably present in all patients with rheumatoid arthritis. Anemia represents the end result of the more severe degrees of the abnormality. Chronic blood loss, hemolysis, increased blood volume, and bone marrow depression represent secondary complications superimposed on the underIying abnormality. Esseva trovate que patientes con arthritis rheumatoidee ha un rapide clearance de ferro in le plasma. Lor concentration de ferro sera1 es basse. IIles ha un reducite absorption de ferro ab le vias gastro-intestinal. Iste anormalitates pare esser manifestationes de un anormal metabolismo de ferro le qua1 es probabilemente presente in omne patientes con arthritis rheumatoidee. Anemia representa le resultato terminal del grados plus sever de iste anormalitate. Chronic perdita de sanguine, hemolyse, augmento del volumine de sanguine, e depression del medulla ossee representa complicationes secundari que se superimpone a1 subjacente anormalitate. A NEMIA IS RECOGNIZED as a common manifestatim of rheumatoid arthritis. Short, Bauer, and Reynolds* found approximately 23 per cent of a large series of patients anemic on the basis of red cell count. The mechanism of production of this anemia has been a subject of investigation, and there is considerable difference in opinions expressed by various investigators. The present paper reports the results of the study of iron metabolism in 29 patients with rheumatoid aithritis. METHODS Hemoglobin determination was done by the Oxyhemoglobin method standardized by the iron determination and the 0, capacity.2 The microhematocrit technique was used,3 and Ted blood cell count and white blood cell count were done by an electronic cell counter.4 Sedimentation rate was done using the Cutler method.5 Osmotic fragility of the red blood cells was carried out according to the method of Parpart.6 The rheumatoid factor was meisured by latex particle fixation.7 The electrophoresis of serum protein was carried out by paper electrophoresis.* Platelet count was done with the phase method.9 Serum iron determinations were done according to the method outlined by Schales.10 Unsaturated isron binding capacity was done by the method of Ventura.11 Direct and indirect Coombs tests were performed at 37" C. using commercial antiserum.12 Blood volume determinations were done by Cr51 tagging of red blood cells13 or with Evans blue dye dihtion technique.14 Total circulating hemoglobin was calculated on the basis of the patient's venous hemoglobin and the blood volume with a correction factor of .93.14 From the Departments of Medicine and Pathology of the Bryn Mawr Hospital, Bryn Mawr, Pennsylvania. This study was supported by a grant from the John S . Sharpe Research Foundation of the Bryn Mawr Hospital. Rcquests for reprints should be addressed to F. Douglas Raymond, Jr., M.D., Aryn Muw?* Mcdical Building, B y n Mawr, Pennsylvania. 233 ARTHRITIS AND RHEUMATISM, VOL. 8, NO. 2 (APRIL), 1965 234 RAYMOND, BOWIE, DUGAN Red blood cell survival was determined after autotransfusion with Cr51 tagged red blood cells.15 Four-day determinations of stercobilinogen were carried out following the method described by Watson.16 Iron clearance, circulating red cell iron, and iron incorpoiation were studied with radioactive Fe59 uaing the method described by Huff.17 Oral Fe'o was given to 9 of our patients in the study of iron absorption. Twenty-five , ~ gof. Fe59 as an aqueous solution was used and was given with 6 ounces of fresh orange juice as a reducing agent. Blood wmples were drawn at 30, 60, and 120 minutes. The patients were fasting, and the absorption was that of the maximum found in the samples drawn at these times. Bone marrow examinations of material obtained by sternal puncture were carried out for determination of cellularity and iron stores. Evaluation of iron stores was graded 0 indicating no staining iron, I t indicating a decreased iron storage, 2+ indicating a normal iron storage, and 3+ indicating increased storage following the method used by Beutler.18 RESULTS All of the patients studied had been diagnosed as classical or definite rheumatoid arthritis according to the American Rheumatism Association Classification.19 All patients had active disease at the time of study which was manifest by joint symptoms and elevated sedimentation rate and were divided into 4 groups of disease activity according to a classification based on Lansbury's Criteria19 with some modification (see table 1).No attempt was made to calculate either an articular or a systemic index. Using the clinical classification described, approximately one-third of our patients had a mild degree of activity; and the remaining two-thirds had a more marked degree of activity (see table 2). Twenty-two of our patients had venous hemoglcbin determinations below 12 Gm./100 ml. Of the remaining 7 patients, only 1 was higher than 13 Gm./ 100 ml. We observed a correlation between the degree of rheumatoid activity and the level of hemoglobin concentration. Of the 10 patients who had Grade I and Grade I1 activity, 5 (50 per cent) had a venous hemogloibin concentration above 12 Gm./100 ml. Of the 19 patients with Grades I11 and IV activity, 2 (12 per cent) had venous hemoglobin concentration above 12 Gm./100 ml. The red blood cell of the anemic patients tended to be hypochromic and normocytic. Seventeen patients (58 per cent) had a mean corpuscular henioglobin of 27 p g . or less (see table 3). The analysis of the mean corpuscula~ volume of the anemic patients showed that about 80 per cent fell within the range of 78 to 100 micra3. The leucocyte count fell within the normal range in all except 5 patients. One of these, J. K., suffering from Still's disease, had a count of 42,550/mm3. Platelet counts were either determined or estimated to be normal in all patients; serum bilirubin determinations were done in all but 3 patients and were all within the normal range. Only 1 patient, M. W., had a definitely elevated blood urea nitrogen which was 37 mg. per 100 ml. Both direct and indirect Coomb's tests were done on 18 of the patients, and both tests were found normal in all subjects. Stool examination for occult blood was done on 15 patients. In these, only 2 showed significant amount of occult blood ( + 2 ) . 235 lRON METABOLISM AND RHEUMATOID ARTHRITIS Table 1.-Gradation of Rheumatoid Activity All had a diagnosis of classical or definite rheumatoid arthritis. All had joint symptoms. All had elevated sedimentation rate. Grade I 1. A.M. stiffness with or without jelling Grade 11 1. A.M. stiffness 2. Diminished hand grips and inflammation of small joints of hands 3. Inflammation of one joint only if hands not involved Grade I l l 1. A.M. stiffness 2. Decreased hand grips and inflammation of small joints Qf hands 3. Involvement of one or two other joints provided severity not marked as determined by large and repeated effusion and muscle wasting 4. No fever Grade IV 1. A.M. stiffness 2. Decreased hand grips and inflammation of small joints of hands 3. More than two other joints involved OT if only two are involved, there is marked involvement including large effusion 4. Fever Table 2 Rheumatoid Activity Number of Patients % of Patients Grade I Grade I1 Grade I11 Grade IV 2 8 1 18 7 28 4 63 The analysis of the plasma volume data indicated that 20 of 29 patients fell within the normal range. Four patients had an increase in excess of 10 per cent of normal, and 6 patients had a reduction of more than 10 per cent. In only 1 patient the increased plasma volume appeared to have a bearing on the reduced hemoglobin concentration (8.2 Gm. per 100 ml). Measurement of the total circulating hemoglobin indicated that 19 out of 29 patients had a significant reduction of the total circulating hemoglobin (less than 75 per cent). Red cell survival by tho Cr”l method was carried out in all patients. In 1 patient the test was found unsatisfactory. Of the remaining 28 patients tested, all but 2 had a normal survival. Both of these patients had splenomegaly and a significant increase in the hemolytic index. This was considered as evidence of a hemolytic component contributing to the anemia. Bone marrow smears and sections were obtained for evaluation of iron stores and evidence of erythroblastic activity. Twenty cf the 25 patients studied had normal or increased iron stores; all 25 patients had either a normal or increased erythroblastic activity. Nine of the patients were given Fe5Qorally as ferrous sulfate for determina- J. E. H. M. L. N. T. D. E. C. H. A. W. B. M. W. L. hil. E. Cr. c. w. M. R. -4. P. A. K. B. J. H. P. R. S. B. M. G. E. D. M. Wer. E. S. T. 0.R. S . M. A. S . R. J. G. B. B. J. K. L.B. M. B. Patient 1 1 2 2 2 2 2 2 2 2 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 Grade of Clinical Activity 8.8 13.5 12.3 11.9 8.9 12.8 11.5 12.8 12.3 10.8 8.2 10.0 10.1 7.0 10.7 10.6 10.4 6.4 10.0 10.4 11.7 9.1 10.6 10.0 12.1 11.1 12.3 11.0 10.3 Hemoglobin (Gm./100 ml.) 21 31 32 28 27 32 24 30 32 27 23 30 25 16 31 25 22 26 26 30 31 21 24 28 25 29 27 24 32 Mean Corpuscular Hemoglobin 109 79 101 116 90 94 100 91 99 93 125 97 108 104 83 107 91 105 106 86 82 106 127 102 78 85 104 120 83 Plasma Volume (% Expected) -- 3+ 0 212+ 2+ 2+ 0 3+ 3+ 0 2+ 2+ 2+ 2+ 2+ 3+ 2f 3+ 2+ 0 2+ 3+ 2+ 0 2+ lron Bone Marrow Cellularity 0.2 1.3 0.2 0.9 1.6 0.1 0.2 0.2 3.5 Oral Iron Absorption (0.3-1.5 70) 20 51 59 68 37 69 40 62 76 60 20 32 31 21 51 51 20 28 59 50 62 26 46 86 33 72 40 14 54 Serum Iron 54 54 42 25 53 55 53 59 33 20 31 38 18 119 51 69 68 34 41 36 26 44 Plasma Fe 51 Clearance Tl/z 32.3 42.3 37.6 35.5 28.2 40.1 36.8 30.7 34.4 36.9 28.8 34.6 33.2 25.1 35.3 35.5 34.4 21.2 33.0 33.4 37.5 31.2 33.0 32.2 38.9 34.9 40.5 36.8 33.1 51.R. _ ~ _ _ _ J. E. H. M. L. N. T. D. E. C. H. A. W. B. M. W. L. M. E. Cr. c. w. H. P. R. s. B. hl. G. E. U. M. Wer. E. S. T. 0.R. S. h4. A. S. R. J. G. B. B. J. K. L. B. M. B. A. P. A. K. B. J. Hematocrit Patient (XZ) 0.3 1.6 2.4 0.3 2.1 0.7 0.4 0.9 0.1 1.5 1.1 4.9 O.s 3.1 1.0 1.6 0.5 0.7 4.0 0.6 1.2 0.9 0.9 0.2 0.4 1.3 1.6 0.1 2.8 Retic. Count 97 95 94 94 94 93 93 73 95 96 90 93 91 91 94 94 95 94 93 94 91 91 94 92 94 93 97 97 (92-96oJo) 24 hru. --_ 2.0 2.3 1.7 2.7 2.3 3.0 2.2 2.2 2.2 2.4 2.1 2.3 2.0 2.6 2.1 1.7 2.3 75 72 58 71 75 63 73 70 71 74 75 74 76 73 76 79 71 7s 78 77 76 82 (1.94-2.46%) loss/day 2.4 1.8 2.6 2.2 2.0 1.8 1.6 2.0 1.6 2.3 1.7 72 77 71 71 73 78 (71-77cJo) 10 days _-____ C F RBC Survival 59 33 47 52 34 59 59 232 13 43 125 195 11 82 61 115 27 41 75 48 36 29 73 46 88 45 mpm,/day (70-162) 13 15 32 59 7 15 60 57 4 26 26 22 11 14 14 11 7.2 9 15 16 10 14 20 18 21 14 Hemolytic Index (11-24) Stereohilinogen 129 322 257 156 47 320 101 220 221 184 257 239 211 175 312 322 633 175 477 276 147 175 184 46 Unsat. Fe Binding Cap. neg., x2 neg., x4 neg., x2 neg., x4 neg., x4 4 day l+ neg., x 3 neg., x4 4 day I+ neg., x4 neg., x3, I f , tr. 1-t,2+ neg., 4+, 1+, tr. neg., x2, I+, 2+ Occult Blood Stool Spec. ( 4 ) 238 RAYMOND, BOWIE, DUGAN tion of iron absorption. Of these patients, 3 fell in the normal range (0.3 to 1.5 per cent) for the dose given. One patient was in the range above normal and 5 were below the normal range. The venous hemoglobin level of all patients in whom the oral absorption of iron was studied was below 12 Gm./100 ml., except for one who had a level of 12.3 Gm./100 ml. Red blood cell iron incorporation was measured in 22 patients using radioactive Fe6ginjected intravenously. Only 2 patients appeared to have reduced incorporation. One of these was the patient with Still's disease, and the other (M. B.) was one of the patients with evidence of hemolysis. Incorporation of Fe59following oral FejQingestion was likewise very poor in this latter patient as well as in the second patient (B. B.) who had evidence of a hemolytic process. All of the patients had repeatedly IQWserum iron levels. These determinations were done fasting at approximately 9 a.m. in order to minimize the diurnal variation. Twenty-seven (93 per cent) of the patients had serum iron determinations below 70 &lo0 ml. and 12 (41 per cent) of the subjects were below 40 @100 ml. W e found some correlation between the clinical activity of the patients and the serum iron levels. Of the group of patients with Grades I and 11 activity, only 2 patients (20 per cent) had serum iron levels below 40 &lo0 ml. Of the patients with Grades 111 and IV activity, 9 patients (47 per cent) had serum Fe5glevels below 40 Pg/lOO ml. Unsaturated iron binding capacity was done on 25 patients. Ninety-two per cent of these had levels which were either normal or decreased. Analysis of electrophoretic patterns indicated that the patients had depressed albumin and elevated alpha2 and gamma globulin fractions. The beta globulin fraction was normal. It is in this fraction that siderophilin is found. Our data indicated no evidence of an abnormality of iron transport. Plasma iron clearance was carried out on 21 patients. The T% of normal individuals in our laboratory is 70 to 120 minutes. In all but 1patient the plasma clearance was less than 70 minutes. The 1 normal clearance occurred in a patient (R. s.) who had a normal hemoglobin, hematocrit, mean c o ~ u s c u l a r hemoglobin, and serum iron. There appeared to be some correlation between the speed of clearance and the serum iron levels. Twelve patients (57 per cent of the patients studied) had T1/2 plasma iron clearance of less than 45 minutes. Of this group, 8 patients had serum iron lei els below 40 rg/100 ml. All the patients in this group were females. The 4 remaining patients had serum iron values of 60, 51, 51, and 42 mg. per 100 rnl. On the other hand, of the group of 9 patients whose clearance was greater than 45 minutes, only 1 had a serum iron level of less than 40 @lo0 ml. Five of the patients were given 700 mg. of saccharated iron intravenously following their initial studies. First follow-up was done after six weeks on 2 of the patients, on 2 after 12 weeks, and the fifth patient was restudied 20 weeks following the intravenous administration of iron. Three of the 5 patients had a significant increase in their hemoglo~binand hematocrit on their first follow-up study. An additional patient had some evidence of improvement but IRON METABOLISM AND RHEUMATOlD ARTHRITIS 239 the follow-up was not adequate. The remaining patient had a steady rise in his values over an eight month period. The follow-up serum iron determinations, however, did not reflect the improvement seen in the hemoglobin and hematocrit and remained generally below normal. In 4 of the 5 cases, the mean corpuscular hemoglobin increased slightly. Over a period of several years, 5 of our patients who were being followed showed a gradual lessening of their rheumatoid activity. In none of these was it thought that a total remission had occurred but 3 of these patients went from Grade IV to Grade I, 1 patient went from Grade I11 to Grade I, and 1 patient went from Grade 11 to Grade I. In the 4 patients going from Grades I11 or IV to Grade I, there was a significant rise in both hemoglobin and hematocrit. In the 1 patienit going from Grade I1 to Grade I, this rise could not be seen. In none of the 5 patients was there a significant change in the serum iron concentrations. All continued to run abnormally low serum iron levels despite the fact that 4 out of 5 had ceased to be anemic based on a hemoglobin of 12 Gm. DISCUSSION In 1943 Robinson2" first suggested that the anemia of rheumatoid arthritis might be the result of an increased plasma volume. Dixon,21 in 1955, using Evans blue dye concluded that increased plasma volume was probably the important mechanism. Weinstein22fcund some increase in blood volume using Cr51 tagged red blood cells but was reluctant to attribute the anemia to this mechanism. Jef€ery,23in 1952, studied 50 patients with rheumatoid arthritis and found plasma volume to be normal using Evans blue dye dilution method. W e have found no evidence that increased plasma volume is an important factor in the etiology of the anemia. Only 4 of our patients had an increased plasma volume and the rest were either normal or had somewhat diminished volume. The possibility that the anemia of rheumatoid arthritis is the result of a reduction in the red cell survival has been postulated by several workers. W e i n s t e i P found a reduction in red cell survival using Cr51 tagged red celIs but felt that the magnitude of the destruction was nolt great enough to account for the anemia. Ebaugh, et al.24 reported normal survival times as well as normal fecal urobilinogen output of rheumatoid patients. Normal fecal urobilinogen studies have been reported also by Jeffrey2"and by F r e i r e i ~ h Red .~~ blood cell survival and hemolytic index were normal in our patients, ruling out hemolytic anemia as an important factor except in 2 patients in whom an hemolytic component was definitely found. There was no evidence that diminished red blood cell production resulting from bone marrow depression played an important role in the production of the anemia in these patients. Chronic blood loss has been considered as a major factor of the anemia of these patients. Douthwaite and L i n t o t P showed by gastroscopic study that aspirin caused localized hyperemia and erosions of normal gastric mucosa. LangeZ7reported a high incidence of occult blood in the stools of normal 240 RAYMOND, BOWIE, DUGAN 100, 2 4 HOUR5 I: U Y 80 U U F 2? 4 V E R R Q E LOSS PfR DAY = 2 14% V HE?IOLYTICL N D W = 21.9. NoICMnL H E M O ( l f O t 3 l N TURUOVER - II - 24 115.4 O R Y S . N O R U R L 9 5 -173 D A V S Fig. 1.-Survival of Cr51 tagged red cells per individual case. . -. * . : L . n e . K m I DRYS Fig. 2.-Red cells ntilization of intravenous ferrous citrate per cent dose of Feb9 citrate in red cells. people taking aspirin by mouth. Stubbi?sxreported the presence of occult blood in 70 per cent of 180 people after aspirin. On the other hand, Scott, et al.2g studied 96 patients for the presence of fecal occult blood using Chromium tagged recl blood cells. Sixty-one G f these patients had rheumatoid arthritis and the studies were carried out during oral administration of salicylates. While 70 per cent of their patients showed some evidence of blood loss after aspirin administration, the average loss was quite small. In only 6 of 64 patients studied did they ccnsider that anemia had resulted from this cause. All of our patients were taking salicylates at the time of study; and based on Scott's work, we must assume that some blood loss occurred. Our tests for occult blood, however, indicated that this was not a serious loss in any patient. Qur studies do not indicate that iron deficiency anemia is a prime factor in the anemia of our patients. Patients with iron deficiency anemia tend to have increased iron absorption and either oral or intralTenous iron will completely correct the deficiency. Ross,~"in 1950, gave iron by mouth to rheumatoid patients and found no appreciable effect on the anemia after 1 month, IRON METABOLISM AND RHEUMATOID ARTHRITIS 241 Jefirey,23in 1952, found that only a small percentage of his patients responded to oral iron and felt that this represented incidental iron deficiency associated with their underlying disease. Our studies on 9 patients with oral radioactive Fefi9indicated a somewhat diminished ability to absorb iron through the gastrointestinal tract. This data is the opposite to what is seen in iron deficiency anemia where oral Fe”Qstudies show an increased gastrointestinal absorption.22 Plasma iron clearance in iron deficiency anemia is increased but only mildly as compared with normals.lx In our patients, plasma iron clearance was very rapid in the majority of patients. Patients with iron deficiency have absent iron storeslXwhile the majority of our patients had normal or increased stores as determined by bone marrow microscopic examination. In iron deficiency anemia, there IS an increased iron binding capacitylR while rheumatoid patients have a normal or decreased iron binding capacity. Bone marrow examination was either normal or showed erythroid hyperplasia in the great malority of our patients. It may well be that same patients with rheumatoid arthritis who are anemic have some increase in their blood volume, some iron deficiency, a hemolytic process, or a diminished red cell production on the basis of bone marrow depression. Any of these factors may occur in these patients complicating their underlying disease state. W e do not believe, however, that these factors either singly or in combination are primarily responsible for the anemia of rheumatoid arthritis. Patients with rheumatoid arthritis appear to have abnormal iron metabolism as a manifestation of their disease. It is reflected in a rapid plasma clearance of iron and a low serum iron. This abnormal iron metabolism is probably present in all patients with the disease and is further evidence that rheumatoid arthritis is a widespread mesenchymal disorder. The exact nature of the metabolic abnormality cannot be deduced from our study. Iron incorporation studies, iron transport, iron stores, and bone marrow production all appear to be normal. The apparent defect in iron absorption through the gastrointestinal tract may well be a reflection of the basic metabolic abnormality. Intravenous saccharated iron brought about some improvement in the hemoglobin and hematocrit levels in our patients but no improvement in the serum iron levels. In the amounts given, correction of an iron deficiency, per se, should have been complete and we believe that there is no evidence that intravenous iron influenced that underlying defect. The degree of abnormality of iron metabolism can be related to the estimation of the clinical activity of the disease. Our patients with marked activity of their rheumatoid arthritis tended to have a lower serum iron than those with a mild clinical activity. Those with the lowest serum iron levels had the most rapid clearance rates. Our studies indicate that even though clinical activity may diminish, serum iron levels remain depressed reflecting the continued presence of an abnormal iron metabolism. Low serum iron levels were found in our patients despite a return of the hemoglobin and hematocrit to normal levels and a marked improvelment in clinical symptoms so that the patients had little more than morning stiffness, occasional joint aches, and elevated sedimentation rate. 242 RAYMOND, BOWIE, DUCAN SUMMARY The anemia that occurs in patients with rheumatoid arthritis appears to be primarily the result of an abnormal iron metabolism. All patients who have the disease for any length of time probably have this metabolic abnormality. Anemia represents the end result of the more severe degrees of the abnormality and occurs in only a minority of patients with rheumatoid arthritis. The abnormality in iron metabolism is manifest by a rapid iron clearance, a low serum iron and perhaps by a diminished ability of the intestinlal wall to absorb iron. Chronic blood loss, hemolysis, increased blood volume, or bone marrow depression may contribute to a reduction of the relative hemoglobin value of circulating blood in patients with rheumatoid arthritis who are anemic. There is no evidence, however, that these are primary factors in the majority of such patients, and when present represent secondary complications superimposed on the underlying abnormality. ACKNOWLEDGMENT We acknowledge the assistance of- Max M. Strumia, M.D., for advice and help in this paper. REFERENCES 1. Short, C. L., Bauer, W., and Reynolds, W. E.: Rheumatoid Arthritis. Cambridge, Harvard University Press, 1957, p. 349. 2. Drabkin, D.: The standardization of hemoglobin measurement. Am. J. Med. Sci. 217:710; 1949. 3. Strumia, M. M., Sample, A. B., and Hart, E. D.: An improved micro heniatocrit method. Am. J. Clin. Path. 249, 1954. 4. Coulter, W. H.: Coulter Electronics, Chicago: High Speed Automatic Blood Cell Counter and Cell Size Analyzer, presented to National Electronics Conference, Chicago, 1956. 5. Cutler, J. W.: A standardized technique for sedimentation rate. J. Lab. Clin. Med. 26:542, 1940. 6. Parpart, A. K., Lorenz, P. B., Parpart, E. R., Gregg, J. R., and Chase, A. M.: The osmotic resistance of human red cells. J, Clin. Invest. 2G:G36, 1947. 7 . Singer, J. M., Plotz, C. M.: The latex fixation test. Am. J. Med. 21:888, 1956. 8. Gudaitis, A. V.: Plasma protein fractionation by paper electrophoresis. J. Kentucky Med. Assn. 57:807, 1959. 9. Brecker, G., and Cronkite, E. P.: Morphology and enumeration of human blood plateIets. J. Appl. Physiol. 3:365, 1950. 10. Schales, 0.: Standard methods of clinical chemistry. Personal Communication. 11. Ventura, S.: Determination of the unsaturated iron-binding capacity of serum. J. Clin. Path. 5:271, 1952. 12. Coombs, R. A., Mourant, A. E., and Race, R. R.: A new test for the detection of weak and incomplete rh agglutinins. Brit. J. Exp. Path. 36: 225, 1945. 13. Gray, S. J., Stelrling, K.: The tagging of red cells and plasma protein with radioactive chromium. J. Clin. Invest. 29:1604, 1950. 14. Keith, N. M., Rowntree, L. G., and Geraghty, J. T.: A method for the determination of plasma and blood volume. Arch. Intern. Med. 18547, 1915 (modified in our lab). 15. Strumia, M. M., Taylor, L., Sample, A. B., Colwell, L. S., and Dugan, A.: Uses and limitations of survival studies of erythrocytes tagged with Cr". Blood 10:429, 1955. 16. Watson, C. J.: The bile pigments. New IRON METABOLISM AND RHEUMATOID ARTHRITIS Eng. J. Med. 227:665, 705, 1942. 17. Huff, R. L., Hennessy, T. G., Austin, R. E., Carcia, J. F., Roberts, R. M., and Lawrence, J. H.: Plasma and red cell iron turnover in normal subjects and in patients having various hematopoietic disorders. J. Clin. Invest. 29: 1041, 1950. 18. Beutler, E., Robson, B. S., and Buttenwieser, E.: A comparison of the plasma iron, iron-binding capacity, sternal marrow iron, and other methods in the clinical evaluation of iron stores. Ann. Intern. Med. 48:60, 1958. 19. Hollander, J. L., and Collaborators: Arthritis and Allied Conditions, 6th Ed. Philadelphia, Lea and Fibiger, 1960, p. 234. 20. Robinson, G. L.: A study of liver function and plasma volume in chronic rheumatism by means of phenoltetrabromphthalein sodium sulphonate. Ann. Rheum. Dis. 3:207, 1943. 21. Dixon, A. St. J., Ramcharan, S., and Ropes, M. W.: Rheumatoid arthritis: dye retention studies and comparison of dye and radioactivity labeled red cell methods for measurement of blood volume. Ann. Rheum. Dis. 14: 51, 1955. 22. Weinstein, I. M.: A correlative study of the erythrokinetics and disturbances in iron metabolism associated with the anemia of rheumatoid arthritis. 243 Blood 14:950, 1959. 23. Jeffrey, M. R.: Anemia of rheumatoid arthritis. Ann. Rheum. Dis. 11:162, 1952. 24. Ebaugh, R. G., Peterson, R. E., Rodman, G. P., and Bunim, J. J.: Symposium on rheumatic diseases. Med. Clin. N. Amer. 39:489, 1955. 25. Freireich, F. J., Ross, J. F., Bayles, T. B., Emerson, C. P., and Finch, S. C.: Mechanism of anemia associated with rheumatoid arthritis. Ann. Rheum. Dis. 13:365, 1954. 26. Douthwaite, A. H., and Lintott, G. A. M.: Gastroscopic observations of the effect of aspirin and certain other substances on the stomach. Lancet 2: 1222, 1938. 27. Lange, H. F.: Salicylates and gastric hemorrhage. Gastroenterology 33:770, 1957. 28. Stubbe, L. T. F.: Occult blood in feces after administration of aspirin. Brit. M. J.. 2:1062, 1958. 29. Scott, J. R., Porter, I. H., Lewis, S. M., and Dixon, A. St. J.: Studies of gastrointestinal bleeding caused by corticosteroids, salicylates, and other analgesics. Quart. J. M. 30:167, 1961. 30. Ross, D. N.: Oral and intravenous iron therapy in the anemia of rheumatoid arthritis. Ann. Rheum. Dis. 9:358, 1950. F. Douglas Raymond, Jr., M.D., Assistant Attending Physician, Service of Arthritis, Department of Medicine, Bryn Mawr H,ospital, Bryn Mawr, Pennsylvania; Instructor in Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania. Mofris A. Bowie, M.D., Attending Physician, Service of Arthritis, Department of Medicine, Bryn Mawi Hospital, B y n Mawr, Pennsylvania; Assistant Professor of Physical Medicine and Rehabilitation, Graduate School of Medicine, University of Pennsylvania; Assistant Professor of Clinical Medicine, University of Pennsylsania School of Medicine, Philadelphia, Pennsylvania . Ann Dugan, B .S., Research Assistant, John S. S h r p e Research Foundation of the Bryn M a w Hospital, Bryn Mawr, Pa.